다중 에이전트의 의사소통 모델에 관한 연구 - KQML 언어의 의미론을 중심으로 - 인지과학 협동과정 98132-502 김 경수.

Slides:



Advertisements
Similar presentations
2 Introduction A central issue in supporting interoperability is achieving type compatibility. Type compatibility allows (a) entities developed by various.
Advertisements

FIPA Interaction Protocol. Request Interaction Protocol Summary –Request Interaction Protocol allows one agent to request another to perform some action.
Web Service Ahmed Gamal Ahmed Nile University Bioinformatics Group
Chapter 16: Multiagent Systems Service-Oriented Computing: Semantics, Processes, Agents – Munindar P. Singh and Michael N. Huhns, Wiley, 2005.
Component Oriented Programming 1 Chapter 2 Theory of Components.
Chapter 16: Multiagent Systems Service-Oriented Computing: Semantics, Processes, Agents – Munindar P. Singh and Michael N. Huhns, Wiley, 2005.
OASIS Reference Model for Service Oriented Architecture 1.0
1 Introduction to XML. XML eXtensible implies that users define tag content Markup implies it is a coded document Language implies it is a metalanguage.
Specifying Agent Interaction Protocols with AUML and OCL COSC 6341 Project Presentation Alexei Lapouchnian November 29, 2000.
Understand Web Services
Software Agent -communication-. Outline Overview Speech act theory Agent communication languages Summary 1/35.
14 -1 Lecture 14 Multi-Agent Systems Topics –Basics –MAS Architectures –Coordination KQML –Collaboration CNP –Development of MAS –Applications.
© 2001 Franz J. Kurfess Knowledge Processing 1 CPE/CSC 580: Knowledge Management Dr. Franz J. Kurfess Computer Science Department Cal Poly.
Chapter 8 The nature of communication
INTERACTION AND COMMUNICATION. Coordination A property of interaction among a set of agents performing some activity in a shared state. The degree of.
Artificial Intelligence and Lisp Lecture 4 LiU Course TDDC65 Autumn Semester, 2010
Multiagent Systems and Societies of Agents
Agent Communication Languages CSE5610- Intelligent Software Systems Agent Communication Languages.
Protocols and the TCP/IP Suite
OWL-S: Semantic Markup for Web Services
Agent Communication Language. Agent Coordination Agents communicate in order to achieve better the goals of themselves or of the society Coordination.
Kmi.open.ac.uk Semantic Execution Environments Service Engineering and Execution Barry Norton and Mick Kerrigan.
KQML as an Agent Communication Language Tim Finin, Yannis Labrou, James Mayfield.
Course Instructor: Aisha Azeem
SNMP Simple Network Management Protocol
Agents Communication Languages (ACL) Dumitru Roman Digital Enterprise Research Institute
Protocols and the TCP/IP Suite Chapter 4. Multilayer communication. A series of layers, each built upon the one below it. The purpose of each layer is.
Design Patterns & Component Frameworks CS67041Spring 2002 Open Agent Architecture (OAA) - Part II -Murali Krishnan.
PROGRAMMING LANGUAGES The Study of Programming Languages.
* SRA Division CoopIS- September Extending Multi-Agent Cooperation by Overhearing Floriano Zini SRA Division - ITC-irst, Povo (Trento) - Italy Joint.
2APL A Practical Agent Programming Language March 6, 2007 Cathy Yen.
Agent Communication The nature of communication Indirect communication
Protocols and the TCP/IP Suite
Intelligent Agents Russell and Norvig: AI: A Modern Approach
Web Services Glossary Summary of Holger Lausen
SIF8072 Distributed Artificial Intelligence and Intelligent Agents 13 February 2003 Lecture 5: Agent Communication Lecturer:
 The design of the imperative languages is based directly on the von Neumann architecture  Efficiency is the primary concern  Low-level specifications:
1 MFI-5: Metamodel for Process models registration HE Keqing, WANG Chong State Key Lab. Of Software Engineering, Wuhan University
EEL 5937 Agent communication EEL 5937 Multi Agent Systems Lecture 10, Feb. 6, 2003 Lotzi Bölöni.
© Enn Tyugu1 Algorithms of Artificial Intelligence Lecture 9: Intelligent agents E. Tyugu Spring 2003.
An Ontological Framework for Web Service Processes By Claus Pahl and Ronan Barrett.
In-Vehicle Networking
Modeling Component-based Software Systems with UML 2.0 George T. Edwards Jaiganesh Balasubramanian Arvind S. Krishna Vanderbilt University Nashville, TN.
Service Creation Model and Framework Focus: control, redirection services Challenges: –Simplicity, user-friendliness –Flexibility –Robustness (e.g., feature.
AOSE Multi-Agent Interaction. Agents and Interaction Interaction forms the basis of an agents collaborative problem solving capabilities. –Agents are.
Supporting FIPA Interoperability for Legacy Multi-Agent Systems Christos Georgousopoulos 1 Omer F. Rana 1 ( ) 2.
Agent Communication in Multi Agent Systems. Reference Weiss – Chapter 2 Wooldridge – Chapter 8.
1 The Architecture and Protocol Design for Underwater Network Presented by James.
Intelligent Agent Framework1 From Chapter 7 of Constructing Intelligent Agents with Java.
CHAPTER 4 PROTOCOLS AND THE TCP/IP SUITE Acknowledgement: The Slides Were Provided By Cory Beard, William Stallings For Their Textbook “Wireless Communication.
EEL 5937 Agent communication EEL 5937 Multi Agent Systems Lotzi Bölöni.
Intelligent Agents: Technology and Applications Agent Communications IST 597B Spring 2003 John Yen.
CSCI 465 D ata Communications and Networks Lecture 2 Martin van Bommel CSCI 465 Data Communications & Networks 1.
Routing Information Protocol
Object-Oriented Design Concepts University of Sunderland.
Agent Communication Languages Speech act theory Speech act theory Semantics of languages Semantics of languages KQML KQML FIPA ACL FIPA ACL Comparison.
EEL 5937 Content languages EEL 5937 Multi Agent Systems Lecture 10, Feb. 6, 2003 Lotzi Bölöni.
for SDN-based flow handover in wireless environments Daniel Corujo Carlos Guimarães Rui L. Aguiar
Agent Communication Michael Floyd SYSC 5103 – Software Agents November 13, 2008.
Models for Resources and Management
LECTURE 9: Agent Communication
Web Ontology Language for Service (OWL-S)
Protocols and the TCP/IP Suite
ECEN “Internet Protocols and Modeling”
Communication in Multi-Agent Systems
Semantic Markup for Semantic Web Tools:
Protocols and the TCP/IP Suite
ONTOMERGE Ontology translations by merging ontologies Paper: Ontology Translation on the Semantic Web by Dejing Dou, Drew McDermott and Peishen Qi 2003.
Toward an Ontology-Driven Architectural Framework for B2B E. Kajan, L
Presentation transcript:

다중 에이전트의 의사소통 모델에 관한 연구 - KQML 언어의 의미론을 중심으로 - 인지과학 협동과정 김 경수

Contents l Agent Communication and Language l KQML(Knowledge Query and Manipulation Language) features n A KQML Message n KQML Performatives l Communication Models of KQML Agent l Framework for Semantics l KQML Implementation l Conclusion

Agent Communication and Language l Communication among software agents is an essential property of agency. l Agent communication language allow agents to effectively communicate and exchange knowledge with other agents, despite differences in hardware platforms.

KQML(Knowledge Query and Manipulation Language) l A high-level communication language and protocol l Independent of content syntax and ontology l An extension set of performatives expressing a belief or an attitude toward some information.

3 Layers of KQML l Communication layer n low level, packet oriented, not stream oriented l Message layer n logic of communication n performative or speech act l Content layer n Language and ontology variable -- specified in message layer

A KQML Message l The ASCII representation of Common Lisp Polish-prefix notation n Represents a single performative n with an associated semantics and protocol n with a list of parameters and its value l readable by humans, simple for programs to parse, and transportable by many inter-application messaging platforms

A KQML Message Example l (ask-if:senderA :receiverB :ontologystd-geo :language prolog :content“spouse(adam, eve)”) Performative Parameter Value

KQML Performatives l The language’s primitives that define the permissible actions that agents may attempt in communicating with each other. l There are some reserved performatives l and reserved performative parameters

Reserved KQML Performatives n Basic query performatives –evaluate, ask-if, ask-in, ask-one, ask-all n Multi-response query performatives –stream-in, stream-all n Response performatives –reply, sorry n Generic informational performatives –tell, achieve, cancel, untell, unachieve n Capability-definition performatives –advertise, subscribe, monitor, import, export n Networking performatives –register, unregister, forward, broadcast, route

Communication Models of KQML Agent l Address Routing l Brokering l Recommending A B ask(X) tell(X) F A AB B broker(ask(X)) tell(X) advertise(ask(X)) ask(X) F recommend(ask(X)) reply(B) advertise(ask(X)) ask(X) tell(X)

Framework for Semantics 1.a natural language description of the performative’s intuitive meaning 2. An expression which describes the content of the illocutionary act 3. Preconditions that indicate the necessary state for an agent: Pre(A)

Framwork for Semantics(cont’d) 4. Postconditions that describe the states of both interlocutors after the successful utterance of a performative(sender), and after the receipt and processing of a message(receiver): Post(A) 5. A completion condition for the performative: completion 6. Any explanatory comments that might be helpful

Semantics for KQML(example) l ask-if(A, B, X) 1. A wants to know what B believes regarding the truth status of the content X 2. WANT(A, KNOW(A, S)) where S may be any of BEL(B, X) or ¬BEL(B, X) 3. Pre(A): WANT(A, KNOW(A, S) ^ KNOW(A, INT(B, PROC(B, M) where M is ask-if(A, B, X) Pre(B): INT(B, PROC(B,M)) 4. Post(A): INT(A, KNOW(A, S)) Post(B): KNOW(B, WANT(A, KNOW(A, S))

Semantics for KQML(example) 5. Completion: KNOW(A, S’) where S’ is either BEL(B,X) or ¬ BEL(B,X), but not necessarily the same instantiation of S that appears in Post(A), for example 6. Pre(A) and Pre(B) suggest that a proper advertisement is needed to establish them

KQML Implementation l MarketPlace Application n A java-based Implementation of inter-agent communication about buy-and-sell actions n Consists of a FacilitatorAgent, and several kinds of BuyerAgents and SellerAgents to interact there n Each agent class extends the CIAgent Framework, which is a JavaBean, supporting event-processing capability.

Communication Model of MarketPlace Application B S F recommend_one(A,F,item1) tell(F, B) advertise(S,F,item1) ask_if(B,S,X) make_offer(S,B,Y) or deny(S,B,X) make_offer(B,S,Y’) accept_offer(S,B,Y’) or reject(S,B) tell(B, S, Z)

Communication Model of MarketPlace Application(2)

Communication Model of MarketPlace Application l 개선점 n KQML specification 에 정확히 일치하지 않는다. –ex. advertise, recommend 의 content 는 performative message n 의미론적 통일성 필요 –ex. tell performative 는 ask_if 의 응답이다.

The Revised Communication Model B S F recommend_one(A,F,ask_if(B,_,X) reply(F, B, “S”) advertise(S,F,_,S,X) ask_if(B,S,X) tell(S,B,X) or deny(S,B,X) make_offer(B,S,Y’) accept_offer(S,B,Y’) or reject(S,B) achieve(B, S, Z) request(B,S,make_offer(S,B)) make_offer(S,B,Y) tell(B, S, Z)

The Revised Communication Model(cont’d) l 구현상의 한계 n KQML parsing module 의 부재 n 따라서, 어떤 performative 는 content 로서 다른 performative 를 가져야 하나, 기존의 메시지 클래스는 중첩된 performative message 를 구현할 수 없다.

Conclusion l 에이전트간 의사소통 언어가 다양한 어플리케이션 간에 통용되려면, 더욱 형식적인 의미론적 기술이 필수적 l 의미론적 기술 (description) 은 performance modeling 을 위한 의사소통 모형을 제공할 수 있어야 한다. l 특정 분야에서 이루어지는 의사소통 행위의 화행분석으로부터 유용한 performative 의 형식을 구축해야 한다.